Skip to main content

Full text of "ARPA Computer Network Working Group May 24, 1968 Agenda"

See other formats

28, 1DGS 
B. Boehm (RA:) 
G. Buck (UCSB) 
D. Engclbart/B. Raphael (SRI) 
D. Evans (UTAH) 
L. Kleinrock (UCLA) 
Y. Part (AEO-D) 
L. Robers/B. Wessler (ARPA) 
Plus additional persons on attached attendance list 
A one day meeting of the ARPA Computer Network Working Group met Friday, 
May 24, 1968, in Boelter Hall, UCLA, Los Angeles. In attendance were those 
.persons nmed in the attached list. The agenda covered the six topics 
discussed below. 
1. Review of the network schedule 
it was estimated that the IMP contractor v½ould be chosen in 
approximately July 1968, and that the IMPs would be installed at 
the sites of the four initial network nodes (UCLA, UCSB, UTAH, 
SRI) by approximately April 1$9. 
2. Reviews' of the simulation effort 
A description was given of the network simulator being' written 
by UCLA. The simulator, being ¾,Titen in GPSS,consists of eight 
blocks: Initialization, Traffic Generatiou, Host to I:dP Routine, 
CPU Routine, Communication Channel, Routing', Channel Errors, and 
I[P to Host Data Transfer. At the present time only the Traffic 
Generation block has been completed. Progress has been slowed 
due to the academic year-end work lead. It was estimated that 
the complete simulation might be .first run by mid-June 1968. 
Only after debugging is completed and after the 360/75 is moved, 
is the actual simulator use anticipated - the target time is 
late July 1968. 
------------------------------<page break>-----------------------------
\;cry short dcsct'iptio:s ',,'ore ,given'of each host - 
a Univac 1!08 at UTAI[, %,,vo SDS 9d0s a't SilK,' a:l SDS Sigma 7 
possibly with an IBM 360/75 at UCLA, and an IBM 360/50 at 
UCSB. Both tolotypov,'rito' and Ci'T type terminals are or will 
be in operation at all locations. 
Three :::ajo:' obocctivos œor the ¾,'orkin Oroup's efforts 
were discussed and appeared to be accopzab!o. Those objectives 
\','Ore a) soivillg' thos host so host communica'ion problems 
necessary to achieve a worki::g' network, b) aczivo participation 
in the debugging of the initial network, :;::d c) interactin s with 
the IMP contractor,. throu.g'h the auspices of ARPA, mutually 
monitoring and advising each other. 
Attcnt!o,a needed to be devoted to 'the areas of hos to 
host communication, host-I}.[P co:umunication, and %ie host-!.liP 
hardware interface. The prescP. focus was di.'ccte½l o the 
area of host to host communication, with activizy in the 
latter areas being deferred until the IIP contractor is chosen 
and ready. 
A target time of October 1959, (six months after the assumed 
April 1959 ISLP installation) was arbitrarily chosen for planning' 
purposes. It was decided that the efforts of the g'roup should 
.be directed toward providin, the capability for any user from 
his console to operate a variety of interactive 
at other hosts. hie jot differences were fo:d to exist in the 
character sots, C]llF{ICtOf codes, COf.;:l:dlld stl':ct!'cS, 
in precedeires even a;on.f hosts of the initinl notwo:'k. 
differences applied %o teletypewriter %ermine!s, CRT terminals, 
textual material, and graphical material. 
It was felt, at this timc, it {'ould be a major chiovcment 
if a user, from his console, via the network, could operate any 
interactive program available to network users, at any host. 
Th user would be responsible for executing the log'in, log'out, 
and prog'ram commands required by the remote host. The movement 
of fils between hosts might be very restricted or not possible 
in some cases. The remote storage of user files 'would depend 
upon the capability of the individual hosts. 
5. Assignment of tasks 
Each site will proceed to 
a) select for.analysis one (or more if possible)' "typical" 
interactive system tl%at it plans to offer to network 
------------------------------<page break>-----------------------------
the :[.,',>'.: and ou'pul. "mos.sa_c" 
p'."c2c:' Lo dal with, and 
c,c the bchavi_o_r of this sy,tcm as it v:ou!d appear 
to the nctv,'oz'k user, 
The  .... . , 4 
c,,.p,,a.. on l,l'.gso analyses would be on toxtua material. 
It was deemcol desirable, if possible, fo' those nlyo to 
encompass g'raphical processes too. 
A secondary area each site will examine,as it feels appropriate 
to its own situation, tile fol!owinf¾ topics: 
a) es'imates of manpou'er, ti:ue, and costs required to 
achieve :he level of capability being' assm.ed here, 
b) problems associated with accocrating for hosc time 
use by "" 
u.O ilO work, alld 
c) possible scheclules of time the host could be available 
o network use. 
6. Working' lI'I'3111TC.*P, lOlltS v.'itll Obllel' :ICi '.vO',.']-: 
I was af;recd tha% the ,:/'fo!'t. oœ the: iT:'oup ':.c:'. %o bc 
consciotlsly directed o the dovclol);,,c.:L of tcck::iq;,c,o ;;:d probloat 
solutions t'hat would have :cLv,'ork-v,'idc app!icatio:;.. Effort was 
not to be directed so as to cause the initial nctv.,ork to form : 
closed community. 
The inclusion of the IMP contractor as a member of the 
working group was felt to be desirable. 
Means of assuring adequate communication with other network 
members couldn't be worked out at this time. Present plans call 
for all documents uod or generated by the working group to be 
made available to the Network Information Center (N%C), Access 
to the NIC should be possible for any network member. Contact 
by members of the working group with other network members should 
also occur on a continuing basis. 
------------------------------<page break>-----------------------------
UTAH - Herschel ]fall 
Robert E. Stephenson 
(SOl) 322-S3-10 
( S..:.,c .'.d,i :', 
UCSB - G oi'd o.u Luck 
UC l'& - Leonard Kleinrock 
John Stehura 
U. C .... :., 
3?2320, ocltor 
d7S-9711 x 7202 or x 7212 
Boolter Hall 3732 
David G. Cantor 
RAND - Irwin Greenwald 
,la t h Dept. 
478-9711 x 4484 or x 2229 
1700 Main Street 
Santa .%Ionlea, Calif. 90408 
(213) 393-0411 x 738S 
SRI - J.F. Rullfson 
Kaye Tomlin 
Elmer B. Shapiro 
Stanford Research Institute 
335 Ravenswood Ave. 
Menlo Park, Calif. 94025 
(415) 526-6200 x dl16 
(Samc Address) 
(d15) 26-5200 x SS20 
(same address) 
(415) 326-6200 x 2859 
------------------------------<page break>-----------------------------